Skip to main content
The roadmap is delivery-driven: items are tracked by quarter and marked when shipped. The timeline below summarizes the platform from launch in 2017 to the current focus on distributed AI compute.

​
Timeline

GamerHash and GHX milestones from 2017 GamerHash Miner App launch through 2026 mining shutdown and full pivot to AI compute

​
Where we’ve been

GamerHash launched in November 2017 as a desktop client that paid users for sharing idle GPU time — years before “DePIN” existed as a category, and it has been running ever since. By August 2018 the platform had ~300,000 registered users; today it has more than 820,000. The native token GHX launched at TGE on 31 December 2020, with the first exchange listing on Bithumb in April 2021. Seven burn rounds between December 2021 and June 2022 permanently removed 8.1% of supply. The strategic pivot toward AI compute began as open-source AI models created demand for distributed GPU power outside of crypto mining. Mining was permanently wound down on 31 January 2026 — the platform now monetizes GPU contribution exclusively through AI inference workloads.

​
What shipped recently

Selected highlights from the last 18 months. For full per-quarter delivery logs, see the GamerHash Medium community updates. 2024
  • GamerHash AI App scaled across image generation (Sana, DreamShaper, Stable Diffusion variants) and LLM deployments (Llama 3.1/3.2, Qwen 2.5, Gemma 2, PHI-4, DeepSeek).
  • Earning module entered internal testing.
  • Stats page, device map, full website redesign.
2025
  • Multi-chain staking & LP: Solana joined Ethereum and BNB Chain.
  • Bridge expansion: Ethereum ↔ Solana via Raydium / Jupiter integrations.
  • Public compute API: the GPU network opened to developers through an OpenAI-compatible API, with a playground covering text-to-image, text-to-speech, image-to-text and video-to-text.
  • Video models: image-to-video, text-to-video (LTX-Video, LTX-2.3) running locally on user GPUs.
  • GamerHash Mobile: latest iOS support, USDC payouts.
Early 2026
  • Mining permanently concluded; platform fully on AI compute.
  • LTX-2.3 video generation deployed locally after OpenAI shut down Sora.
  • Workflow-automation integrations shipped; more than 10,000 developers now registered against the network.

​
What’s next

Q3 2026 — a more approachable agent ecosystem. A direction rather than a commitment: we want building and running agents on this network to be markedly easier than it is today. That means familiar, OpenAI-compatible interfaces so existing agent tooling works without a rewrite, better defaults so a useful agent doesn’t start with a configuration exercise, and simpler paths for people who aren’t developers to put agents to work on hardware they already own. Specifics will firm up as the work does. Deeper AI capability — multimodal modules (upload images and PDFs, interact in chat), text-to-speech and voice transcription, custom LoRA and user-uploaded models, image-to-video improvements. Platform foundations — redesigned dashboards, Web3 login, version-management improvements, control-panel refresh.

​
Long-term direction

Beyond the quarterly plan, four bets shape the longer arc:
  • One AI app, all open-source models. A single client that runs the most popular open-source models for image, video, voice, and text — locally on the user’s GPU.
  • Earning while running. GPU contribution monetized while the user is actively using AI tools, removing the trade-off between use AI and earn from AI.
  • Distributed compute network. A global pool of consumer GPUs accessible to AI projects, developers, and enterprises through public APIs.
  • Multi-chain GHX. Continued bridging and utility expansion across Ethereum, BNB Chain, and Solana.

​
Reference